Retirement of classic network for ApsaraDB for MongoDB

Since Alibaba Cloud launched its public cloud services in 2011, the classic network has provided network infrastructure for ECS and served a vast number of customers. In 2014, we introduced Virtual Private Cloud (VPC), a modern network architecture that offers enhanced security, stability, flexibility, and performance. As VPC has become the default choice for most customers, ApsaraDB for MongoDB is retiring all instances that use the classic network.

ApsaraDB for MongoDB stopped allowing the creation of new classic network instances on February 21, 2022. This retirement affects all existing instances that have a classic network configuration. To minimize the impact on your workload, we provide a long transition period for you to plan, test, and migrate to a VPC. We will also provide the necessary tools, documentation, and support to help you through this process.

Timeline and impact

Effective 00:00 on October 30, 2024: Renewals, configuration changes, and billing method changes will be disabled

Renewals, configuration changes, and billing method changes for MongoDB instances that use the classic network are disabled.

Note
This affects all instances configured with a classic network, including instances that use hybrid access, even if your applications do not use the classic network endpoint.
Effective 00:00 on February 28, 2025: Service will be discontinued
- Classic network endpoints will be deleted. If your applications are still using a classic network endpoint, their connections will be dropped and will not be automatically restored.
- If an instance only has a classic network endpoint and your account has an existing VPC, the network connection may be switched to a random VPC when the classic network endpoint is deleted.

Note

The retirement of the classic network will be rolled out in batches. You will be notified of the specific schedule through SMS, email, and internal messages. Until then, you cannot renew or make configuration changes to these instances.
After 00:00 on February 28, 2025, or after the classic network endpoint is retired, affected instances will continue to be billed as usual. If you no longer need an affected instance, you must manually release it.
MongoDB instances that are deployed only in a VPC and do not have a classic network endpoint are not affected.
If you migrate your MongoDB instance from the classic network to a VPC and release the classic network endpoint before the retirement date, the instance will not be affected.

Find affected MongoDB instances

You can find instances that use the classic network in the console or by using OpenAPI:

Find instances in the console
1. Log on to the ApsaraDB for MongoDB console.
2. Based on the instance architecture, click Replica Set Instances or Sharded Cluster Instances in the left-side navigation pane.
3. In the upper-left corner of the page, select the resource group and region where the instance is located.
4. In the instance list, filter by Network Type, select Classic Network, and then click OK.
Find instances by using OpenAPI

Call the DescribeInstances operation with the NetworkType parameter set to CLASSIC to find all MongoDB instances that use the classic network in a specified region.

For MongoDB instances with hybrid access enabled, the Network Type is displayed as VPC. You cannot find these instances by using the preceding methods. Check your internal messages and emails for notifications about affected instances. After you receive a notification, take the appropriate actions in the console to prevent service disruptions.

Migrate from classic network to VPC

Important

The following methods apply only to instances that use local disks. If your instance uses cloud disks, submit a ticket for migration assistance.

ApsaraDB for MongoDB allows you to change the network type of an instance from classic network to VPC. This lets your client and the instance communicate within the same VPC, providing greater security and lower network latency.

For existing instances that use the classic network, you must migrate them to a VPC before the service is discontinued to avoid disruptions to your applications. Two migration options are available:
- Option 1: Switch to a VPC without retaining the classic network endpoint. With this option, the VPC endpoint reuses the original classic network endpoint address, so you do not need to change your client's connection string. For more information, see Switch the network type from classic network to VPC.
  
  Important
  With this option, you do not need to change the client connection string, but the instance will experience a transient disconnection.
- Option 2: Use hybrid access for a smooth migration. This option allows you to migrate from the classic network to a VPC without any transient disconnection or service interruption. For more information, see Smoothly migrate from classic network to VPC by using hybrid access.
  
  Important
  With this option, the instance experiences no transient disconnection, but you must modify the connection string on your client to use the new VPC endpoint.
If your instance is already configured for hybrid access between the classic network and a VPC, release the classic network endpoint after updating your applications to use the VPC endpoint. You can release the classic network endpoint from the console or by using OpenAPI:
- Console: For more information, see Modify the expiration time of a classic network endpoint.
- OpenAPI: Call the ModifyDBInstanceNetExpireTime operation and set the ClassicExpendExpiredDays parameter to 0.

FAQ

Q: I have already migrated to a VPC. Why did I receive an internal message that my classic network endpoint has expired but has not been released?

A: An expired classic network endpoint is not automatically released. You must manually release it in the console. For more information, see Modify the expiration time of a classic network endpoint.
Q: I have already migrated to a VPC. Why can't I renew my instance?

A: You cannot renew an instance that still has a classic network endpoint. You must release the classic network endpoint before you can renew the instance. For more information, see Modify the expiration time of a classic network endpoint.